Cortical Tubers' Transformation in Pediatric Patients Diagnosed with Tuberous Sclerosis Complex: A Retrospective...

Background: Cortical tubers (CTs) are hallmark brain lesions in tuberous sclerosis complex (TSC), historically considered stable in number over time; prior literature has correlated overall CT burden on magnetic resonance imaging (MRI) with disease severity. As longitudinal imaging studies assessing CTs' evolution over time are lacking, we aim to investigate temporal changes in CTs-both in number and signal-on MRI in a cohort of pediatric TSC patients. Methods: A retrospective single-center analysis was conducted on 57 pediatric TSC patients who underwent longitudinal MRI studies in a 10-year span. Required MRI sequences included volumetric unenhanced T1-weighted, SWI, T2w and/or FLAIR. CTs were evaluated by two neuroradiologists and classified into five subtypes (A, B, C1, C2, D) according to signal characteristics. Statistical comparison was performed using t-tests. Results: Paired t-test analysis demonstrated a significant longitudinal increase in the overall number of CTs, rising from 16.11 ± 12.43 at baseline to 18.77 ± 13.29 at follow-up (mean difference = -2.67, 95% CI [-3.94, -1.39]; t (56) = 4.19; p < 0.0001), corresponding to a moderate effect size (Cohen's d ≈ 0.56). When stratified by age, patients <2 years-representing the incompletely myelinated subgroup-showed a more pronounced increase in CT burden, from 19.46 ± 15.21 to 24.17 ± 15.75 (mean difference = -4.71, 95% CI [-7.37, -2.04]; t (23) = 3.65; p = 0.0013; d ≈ 0.75). In contrast, patients aged ≥2 years demonstrated a smaller but still significant increase, from 13.67 ± 9.45 to 14.85 ± 9.64 (mean difference = -1.18, 95% CI [-2.08, -0.28]; t (32) = 2.68; p = 0.0115; d ≈ 0.46). Direct comparison between the two subgroups using Welch's two-sample t-test confirmed that the mean CT count in patients <2 years was significantly higher than in those ≥2 years (mean difference = 3.53 ± 1.36; t = 2.59; df = 28.4; p = 0.0075), with a large effect size (Cohen's d ≈ 0.78). Type C1-C2 tubers evolved from pre-existing earlier-stage lesions, while most newly visible CTs over time were type A-B. Type D tubers remained rare and derived from earlier-stage CTs. Conclusions: Contrary to previous assumptions, CTs in pediatric TSC showed a tendency to increase in number and evolve in signal over time, thus challenging the notion of stability and suggesting dynamic behavior. Incomplete myelination in early infancy may impact MRI CTs detection by reducing contrast with surrounding brain tissue, potentially leading to their underestimation/misidentification.